|
|
In article <3fa05697$1@news.povray.org> , Brent G <pov### [at] bc-hqcom>
wrote:
>> But that doesn't show the memory allocated by the application. It only
>> shows the memory used by the application process. Besides, the user didn't
>> mention the platform POV-Ray was running on.
>
> I was merely pointing out that is IS possible
See, the catch is that there is a big difference between the memory
allocated by the application for realy use and what the system shows as the
whole amount of memory used by the application. It may not be clear to a
user, but for the developer it is clear that what the users can see has not
much to do with the realy amount of memory currently allocated by the
application. As strange as it may sound, the system cannot efficiently
determine how much memory is exactly used by an application.
Thorsten
____________________________________________________
Thorsten Froehlich, Duisburg, Germany
e-mail: tho### [at] trfde
Visit POV-Ray on the web: http://mac.povray.org
Post a reply to this message
|
|
|
|
Christoph Hormann wrote:
>I am pretty sure if the observation is correct you can reduce the amount
>of memory shown by the windows task manager by rendering in several
>parts - i can't think of any advantage of this though.
My thinking was that if my machine has, say, 500MB extra RAM (combining real
and swap space, after the O/S and other stuff is loaded), and an image
takes 600MB RAM total to render, then I'm a bit hosed (w/o a RAM upgrade or
closing other stuff down). But if I could render in two parts that each
take 400MB, then it's possible (with much swapping, perhaps). It's my
understanding that as far as the O/S running out of memory, the value that
top or taskman reports actually does matter, regardless of how that memory
is divvied up inside a process.
Or am I way off base (a distinct possibility)?
Dave
Post a reply to this message
|
|